基于改进MeanShift和SURF的目标跟踪

传统颜色直方图的MeanShift（MS）算法只考虑了目标颜色的统计信息，不包含目标的空间信息，当目标颜色与背景颜色相近或目标对象发生光照变化时，容易导致不准确跟踪或跟踪丢失。针对该问题，提出了一种融合改进MS和SURF的跟踪算法。改进的MS算法根据目标对象的最新外接矩形尺寸，确定对象的分块方法，根据各块的Bhattacharyya系数值，确定各块的权重系数，获得初步的跟踪结果；采用SURF特征匹配和校正的方法对其初步跟踪结果进行调整；采用线性加权的方法融合改进的MS和SURF跟踪结果，得出最终的跟踪结果。实验表明，提出的融合改进Ms和SURF的跟踪算法，比传统的Ms算法和固定分块的MS算法都具有更好的跟踪性能。

Object tracking based on improved Mean Shift and SURF

Traditional color histogram Mean Shift （MS） algorithm only considers object＇ s color statistical information, and ignores object＇ s space information, so when the object color is close to the background color, or the object＇ s illumination changes, the tradi- tional MS algorithm easily causes object＇ s tracking inaccurately or lost. Aiming at this issue, a new tracking algorithm which fuses improved MS and SURF is proposed. The improved MS algorithm gets the preliminary tracking results, which determines block method by the size of the lastest enclosing rectangle and determines their weight coefficient by the Bhattacharyya coefficient of each block. After obtaining the tracking result with MS, this algorithm utilizes SURF to refine it. This algorithm uses linear weighted method to fuse the improved MS＇ s tracking results and SURF＇ s. Experimental results show that the new method which fuses improved MS and SURF is better than the traditional MS algorithm and fixed block MS algorithm in tracking performance.